Friday, January 04, 2013

I am skeptical about bumetanide treatment of autism

Let me first disclaim that, while I can publish enough scientific articles to match my job description, I do not consider myself really worth the honorary qualification of a "scientist". Nevertheless, I often use the opportunity to give my twopence on scientific subjects and on the use of science to make life better.

In the December 2012 issue of Translational Psychiatry, E. Lemonnier et al. published an article titled A randomised controlled trial of bumetanide in the treatment of autism in children. The team claims to have tested the diuretic bumetanide by a double-blind, placebo-controlled clinical trial on a group of 60 autistic children aged 3-11 and to have achieved significant improvements. The presumed mechanism of action involves GABA-ergic neurons.

My intuition says simple stuff given orally is, to say the least, highly unlikely to bring improvement in autism spectrum disorders or other conditions based on fundamentally different wiring of the nervous system. Any study reporting such improvement immediately switches my alarms on. And when I read the actual report, I typically find more alarm-switchers.

In previous years, I have written two posts on what I consider quackery concerning another condition - attention deficit and hyperactivity disorder (ADHD). My 2008 post I am skeptical about food additives - hyperactivity link featured the theories that ADHD was caused by "food additives" (?!) or TV watching. My 2011 post ADHD quackery in scientific journal, again addressed the idea to treat ADHD with restricted elimination diets. The stidies I criticized in these posts, similarly to the bumetanide treatment mentioned above, were all published in respected peer-reviewed scientific journals. Unfortunately, this doesn't automatically guarantee good science, let alone good ethics.

Let me cite these older texts. From the 2008 post: "Why didn't anybody try to conduct a study on animal models? At least, I cannot find such an article in PubMed. Animal studies are generally more standardized and hence more reliable than human ones. I know that in many countries it is easier to obtain a permit to experiment on humans than on animals, but still, why not get to the work seriously and do first the paperwork required and then the animal study itself? Why was the study done only on children, after hyperactivity problems, when present, are thought to persist for life? Is it because adults are generally happy with their own flawed selves but demand perfection from their children, relentlessly drawing the little ones to some superhuman standards of intelligence and behaviour?"

From the 2011 post: "People of science have a saying that extraordinary claims require extraordinary evidence. Any claims for successful treatment of a socially important condition are extraordinary... I would ask again, as I did in my old post, why wasn't the study done first on animal models? And if someone thinks animal models of ADHD are not satisfactory (i.e. fail to produce the crazy results wanted and expected by the researcher), why wasn't the experiment done first on adult volunteers with ADHD? Maybe because no adult, except some patients with much more severe diagnoses than ADHD, would agree to participate in such a study; but parents eager to streamline their disabled or just different children easily fall into the trap of wanting the child "either cured or dead"."

For the record, while none of the crazy hypotheses criticized in these two posts of mine has been specifically and inequivocally disproven, they have not been confirmed, either. The 2011 review by T. E. Froehlich et al. Update on Environmental Risk Factors for Attention-Deficit/Hyperactivity Disorder discusses a total of 26 (!) environmental factors seriously considered as contributing to ADHD. Most of these factors are ubiquitous, e.g. "Prenatal Caffeine Exposure" and "“Western” Dietary Patterns". Anybody with even rudimentary background and understanding of scientific research will immediately figure out that we are still deep in the woods. I would add that the only useful information in the cited review is the brief mentioning of "ADHD heritability estimates of 60% to 80%".

Double-blind, placebo-controlled trials are rightly considered the gold standard of clinical research. However, I am putting the bar for the gold standard higher. To satisfy my requirements, a study must also:

1) Be done on animals first, unless suitable animal models don't exist, and
2) For life-long conditions, include adult subjects, unless there are very strong reasons to believe that the proposed methods would not work for adults.

These two requirements were not met by the recent study on bumetanide treatment of autism. Explaining their hypothesis in the Introduction chapter, the authors wrote, "GABAergic signals are altered in autism as evidenced by the following: The excitation/inhibition ratio is modified in experimental models..." (5 references cited). However, to my best knowledge, the researchers never tried to test their working hypothesis on these models that so wonderfully provided argumentation to start an experiment on human beings. Also, the authors did not recruit autistic adults for their study. They recruited children aged 3-11. If their idea was to interfere with the full development of the autism spectrum disorder, they should have taken children between 1.5 and 4 years. At the age of 11, autism is considered as fully established as at age 20. Why, then, not recruit adults? I'll tell you why. Because it would be difficult to convince adults to volunteer for such a study, while many parents are happy to put their children through it. In other words, the study was based on the worrying tendency of parents of disabled children to make for their children choices that nobody ever makes for himself.


Anonymous said...

Avascov Dear Maya
Please feel free to be skeptical. I too am a huge skeptic, but a double-blind, placebo-controlled trial was good enough for me.
So I have tried bumetanide for 3 weeks with my 9 year old autistic son. The results are great but I remain a huge skeptic of un-tested interventions. I have never given my son any other drugs/supplements for autism (eg Prozac/DMG etc) because there is no serious evidence that they do any good. Bravo to the French for doing some genuine research. Also Bumetanide is cheap like paracetemol.


Maya M said...

Peter, if you are still here, please read my new post: "Experiment on disabled children first, then maybe on animals, if you wish"

Anonymous said...

Bumetanide is an old known drug (primarily it is a diuretic) and the effects on humans are well known. It is easier to get approval for a clinical trial that involves an older drug.
If you want to see studies on animals, then search pubmed articles about Bumetanide from the '80s.
Another aspect: there are only a few animal models of autism and all the diagnostic criteria that we have for clinical autism cannot be used for animals.

Maya M said...

I did not mean an animal study to test the safety of bumetanide - I am sure such studies have been done. I meant an animal study on a model of autism to test the drug efficiency. You are right that it is more difficult to assess autism in experimental animals than in humans. However, the facts that the authors cite animal data as leading to their hypothesis, and are now considering an animal study of the sort I propose, strongly suggest that my idea has some merit.